IITB DATA SCIENCE DATATHON

In this notebook, I have performed exploratory data analysis and visualization using the seaborn and MatplotLib both of which are python libraries. After the preprocessing of the data, I used some off-the-shelf models and trained them on the given training data. But even after analyzing so many models, I finally concluded that a KNeighborsClassifier shows 60% accuracy on the training data, 40% accuracy on the validation data, and only 35% accuracy on the test data. But this is the highest I was able to achieve using the given training data and state-of-the-art models.